Chapter 14

Directed Evolution of Carotenoid Synthases

for the Production of Unnatural Carotenoids
Maiko Furubayashi and Daisuke Umeno

Abstract
Directed evolution is a well-established strategy to confer novel catalytic functions to the enzymes. Thanks
to the relative ease of establishing color screening, carotenogenic enzymes can be rapidly evolved in the
laboratory for novel functions. The combinatorial usages of the evolvants result in the creation of diverse
set of novel, sometimes unnatural carotenoids. This chapter describes the directed evolution of diapophy-
toene (C30 carotenoid) synthase CrtM to function in the foreign C40 pathway, and the use of the CrtM
variants thus obtained for the production of novel backbone structures.

Key words: Directed evolution, Library construction, Error-prone PCR, Color screening,
Combinatorial biosynthesis

1. Introduction

Industrial value of carotenoids ranges from anti-oxidants, natural

pigments, hormonal drugs, and micronutrients. Functions of these
carotenoids are based on their molecular structures (molecular
size, solubility, effective number of conjugated double bonds, and
presence/absence of functional groups or cyclic ends), justifying
the effort to construct biosynthetic pathways for various carote-
noids on demand. Dozens of carotenogenic enzymes have been
functionally expressed in noncarotenogenic hosts, such as
Escherichia coli, and the combinatorial expression of these enzymes
often yields the pathways for novel carotenoids (1). When the
desired functions are unavailable (or do not exist in nature), natu-
ral carotenoid enzymes can be engineered to fulfill them (2).
Phytoene desaturases (3, 4), cyclases (3, 5), hydroxylases (6), and
carotene synthases (7) have been successfully engineered to have
altered functions. These “mutant” enzymes are assembled into the

José-Luis Barredo (ed.), Microbial Carotenoids from Bacteria and Microalgae: Methods and Protocols, Methods in Molecular Biology,
vol. 892, DOI 10.1007/978-1-61779-879-5_14, © Springer Science+Business Media, LLC 2012

245
246 M. Furubayashi and D. Umeno

new pathways (natural or unnatural) to the targeted carotenoids.

Carotenoid biosynthetic enzymes cannot be rationally engineered
due to the lack of structural information. Instead, directed evolu-
tion can be well adapted for the creation of novel carotenoid
biosynthetic enzymes (2, 8).
Directed evolution is divided in two experimental steps: library
creation and functional screening (Fig. 1). Starting from one or a
set of parental genes, a large set of genetic variants is prepared.
Thus, created “enzyme library” is then subjected to the screening
to isolate the clones with novel/desired catalytic functions. Due to
the pigmentation of the pathway products, high-throughput color
screening can be developed to isolate the desired enzymatic func-
tions along the path. For the enzymes, such as desaturases, cycla-
ses, and ketolases, one can often see the emergence of new
functions. This is because these enzymes directly alter the photo-
chemical properties of carotenoids and thereby change in functions
can be visualized. For evolving enzymes catalyzing the invisible
steps in carotenoid synthesis, several enzymes are co-expressed to
visualize their cellular functions. As a case study, this chapter
describes the directed evolution of diapophytoene (C30 carotenoid)
synthase CrtM to function in the foreign C40 pathway. CrtM vari-
ants thus obtained are used for the production of novel backbone
structures, which can be further diversified by the addition of
decoration enzymes.
Before the directed evolution experiment, screening system
needs to be set up: one should construct a carotenoid pathway,
including the target enzymes as its component. The pathway
should be designed so that the change in the target enzyme prop-
erties results in the change in colony hue. E. coli has been an exclu-
sive host for evolving carotenogenic enzymes due to the lack of its
background color, rapid growth, and wealth in the genetics. In the
case of screening for phytoene synthase activity, geranylgeranyl-
diphosphate synthase (CrtE) is expressed to establish the precursor
supply. Because the product of phytoene synthase is colorless,
phytoene desaturase (CrtI) is additionally expressed to convert
phytoene into red pigment, lycopene (see Note 1 and Fig. 2).

Fig. 1. General scheme of the directed evolution of carotenoid enzymes.

 14 Directed Evolution of Carotenoid Synthases… 247

Fig. 2. Screening constructs for the CrtM mutants with phytoene synthase activity.

2. Materials

2.1. Library Creation 1. PCR primers forward and reverse (see Note 2).
(PCR Random 2. pUC-crtM (see Note 3 and Fig. 2) (7).
Mutagenesis)
3. Taq polymerase.
4. 10× Taq buffer.
5. 10× dNTP mixture: 2 mM each of dATP, dTTP, dCTP, dGTP.
Prepare 50 μL aliquots of this mixture (to avoid excessive
freeze/thaw cycles) and store at −20°C.
6. 10× MnCl2 solution: 10 μM prepared in distilled water and
stored at room temperature.
7. Nuclease-free water.
8. pUC-based vector (see Note 3) (7).
9. DNA clean and concentratorTM—5 (Zymo Research Corporation,
Orange, CA, USA).
10. Restriction enzymes XbaI and XhoI.
11. T4 DNA ligase.
12. TAE: 40 mM Tris–acetate, and 1 mM ethylene diamine tetraa-
cetic acid (EDTA).
13. Ethidium bromide: 0.5 μg/mL in TAE.
14. Agarose gel: 0.7% LE agarose in TAE and ethidium bromide.
15. ZymocleanTM Gel DNA Recovery Kit (Zymo Research
Corporation, Orange, CA, USA).

2.2. Color Screening 1. Screening plasmid (for details, see Note 4 and Fig. 2).
2. Z-Competent E. coli Transformation Kit (Zymo Research
Corporation, Orange, CA, USA) (see Note 5).
3. E. coli XL1-Blue (Stratagene, La Jolla, CA, USA) (see Note 5).
248 M. Furubayashi and D. Umeno

4. Bio Trace NT Membrane, 82-mm discs (PALL Gelman

Laboratory, Ann Arbor, MI, USA).
5. Sterilized toothpicks.
6. Carbenicillin 1,000× stock: 50 mg/mL (see Note 6).
7. Chloramphenicol 1,000× stock: 30 mg/mL.
8. LB (Luria–Bertani): Add 10 g of bacto-tryptone, 5 g of yeast
extract, and 5 g of NaCl into 1 L of deionized water. Autoclave
at 121°C for 15 min (9).
9. LB-agar: LB and 20 g/L agar.
10. SOC: Add 20 g of bacto-tryptone, 5 g of yeast extract, and
0.5 g of NaCl into 990 mL of deionized water. Add 10 mL of
a 250 mM solution of KCl. Autoclave at 121°C for 15 min.
Just before use, add 5 mL of a sterile solution of 2 M MgCl2
and 20 mL of a sterile 1 M solution of glucose (9).

3. Methods

3.1. Library Creation There are a variety of established methods for the construction of
an enzyme library (10). Among them, whole gene error-prone
PCR is probably the most popular method. Addition of the Mn2+
into the PCR sample lowers the fidelity of the polymerase whereby
inserting random base substitutions into the target sequence (11).
1. For each PCR sample, add to tube: 2 fmol template DNA (see
Notes 7 and 8), 5 μL 10× Taq buffer, 5 μL 10× dNTP mixture,
5 μL of 5 μM each primers (to be 25 pmol in final concentra-
tion), 1 μL Taq polymerase (5 units), 5 μL 10× MnCl2 solu-
tion (to be 10–50 μM in final concentration, see Note 9), and
distilled water to a final volume of 50 μL.
2. Mix the sample by pipetting, and confirm all the solution is on
the bottom.
3. Place the tubes in a thermal cycler, and run the PCR program
below: (1) 5 min initial denature at 94°C. (2) 30 s denature at
94°C, 30 s annealing at the temperature designed for primers,
1 min extension at 72°C. Repeat for 25 cycles. (3) 10 min at
72°C for final extension.
4. Run the product for gel electrophoresis to estimate the yield of
the full-length gene (see Notes 7–9).
5. Clean the PCR product by using a Zymo DNA clean and con-
centrator Kit. In parallel, clean up the vector in the same way.
6. Digest both PCR product and the expression vector with
appropriate enzymes (XbaI and XhoI).
 14 Directed Evolution of Carotenoid Synthases… 249

7. Gel-purify the digested samples using ZymocleanTM Gel DNA

Recovery Kit (see Note 10).
8. Treat the mixture of the purified insert (100 ng) and vector
(100 ng) with T4 DNA ligase for 2 h (400 units/reaction to
total volume of 10 μL) (see Note 11).
9. Optionally, after the reaction, concentrate the ligation product
by using a Zymo DNA clean and concentrator kit (see
Note 12).

3.2. Color Screening 1. To transform the ligation product into E. coli XL1-Blue
Z-competent cells harboring the screening plasmid (see
Note 5), add 5 μL of the ligation product into 100 μL compe-
tent cell on ice.
2. Incubate on ice for 10 min.
3. Add 400 μL SOC medium and shake for 1 h at 37°C.
4. Meanwhile, prepare the screening plates: mount the nitrocel-
lulose membrane on each of 5–10 LB agar plates containing
chloramphenicol and carbenicillin. Handle the nitrocellulose
membrane with tweezers, not with bare hands. The nitrocel-
lulose membrane is used to provide a white background to
facilitate visualization of colony color.
5. Spread all the transformants on the screening plates (see Notes
13 and 14).
6. Incubate the plates at 37°C until the colonies become visible in
size (see Note 15).
7. Leave the plates at room temperature for additional 12–36 h
until the color develops (see Note 16).
8. When the color of colonies became distinguishable, determine
which ones should be the positive variants (see Notes 17–19).
9. Pick all of the pink (positive) clones using sterile toothpicks and
inoculate into a 2 mL LB liquid media containing chloramphen-
icol and carbenicillin. Shake the culture overnight at 37°C.
10. Miniprep the culture to collect the mutant plasmids.
11. At this point, plasmid contains both the mutant plasmid and
the screening plasmid. To remove the latter, retransform the
plasmid into E. coli strain (cloning strains without plasmids).
Spread it on an LB agar plate containing carbenicillin. Pick the
colony with no pigmentation, culture in LB liquid media, and
miniprep to obtain the pure mutant plasmid.
12. For routine analysis of the mutants, determine the genotype of
the mutant genes by DNA sequencing. Cellular function of
the variants can be analyzed by the co-expression of various
carotenogenic genes, followed by the pigment extraction and
product analysis using HPLC-MS.
250 M. Furubayashi and D. Umeno

3.3. Toward the Novel Sequence analysis of the above-obtained CrtM mutants revealed
Carotenoids two amino acid residues (F26, W38) that control the size specificity
of this enzyme (12, 13). In a different directed evolution experi-
ment, farnesyl diphosphate synthase (FDS) from Bacillus stearo-
thermophilus was also engineered into geranylfarnesyl diphosphate
(C25PP) synthase (14). Co-expression of this FDS mutant with
size mutants of CrtMs resulted in the production of the novel
phytoene-type carotenoid with C45 (C25 + C20) and C50 (C25 + C25)
backbones (7). Further addition of carotenoid desaturases CrtI
resulted in the production of diverse set of C45 and/or C50 carote-
noids without protein engineering (15). Many of the downstream
enzymes were shown to accept various different carotenoids. This
way, structural variation of the carotenoids can be easily expanded
by the addition of laboratory-evolved carotenogenic enzymes in a
toolbox of the combinatorial biosynthesis.

4. Notes

1. For the convenience in mutant isolation and subsequent

analysis, we prefer to place the target gene (the random library
of crtM) on a pUC-based high copy expression vector (Fig. 2).
The additional genes (crtE and crtI) are grouped into another
pACYC-based vector. By physically separating the library (in
expression vector) from other genes in the pathway (grouped
in screening plasmid), one can significantly reduce the false
positive clones by the accidental incorporation of mutations in
nontarget genes during DNA handing. Note that this two-
plasmid system has some drawbacks: first, one must prepare
the competent cells harboring the screening plasmid by his/
her own (for this purpose, we recommend Z-Competent
E. coli Transformation Kit; see Note 5). Secondly, every time
after the color screening, the screening plasmids must be
removed from the miniprep of the positive colonies (see
Subheading 3.2, step 11).
2. PCR primers are designed to share the same or similar melting
temperatures. They are used at 5 μM and stored at −20°C. The
primers for error-prone PCR are designed to anneal to the out-
side of the reading frame of the gene so that the entire reading
frame are randomized. In this case, however, regulator sequence
(promoter and ribosome binding site) is also under mutagen-
esis. In the situation where the mutations in the regulation
sequence could lead to the false-positives, design the forward/
reverse primer to anneal to the N/C terminus of the reading
 14 Directed Evolution of Carotenoid Synthases… 251

frame, respectively. In this case, the priming sites of the target

genes are virtually “masked” from the mutagenesis.
3. crtM gene is placed under lac promoter (see Note 20) followed
by optimized ribosome binding site (AGGAGG). crtM library
is to be ligated into the XbaI–XhoI site of the plasmid.
4. Two genes (crtE, crtI) are placed under lac promoter (see
Note 20) followed by optimized ribosome binding site
(AGGAGG).
5. Z-competent cells harboring screening plasmids. Z-Competent
E. coli Transformation Kit is highly recommended to make
competent cells of your own. Following the simple protocol,
the transformation efficiency reliably exceeds 108
transformants/μg plasmid. With this efficiency, >104 colonies
(variants) can be screened for single transformation using 1/5
of the ligation product.
6. The use of carbenicillin instead of ampicillin is important to
prevent the formation of satellite colonies. Screening some-
times requires 2–3 days to fully develop the colony colors, and
ampicillin does not stay effective that long.
7. In the given (fixed) concentration of Mn2+, the average muta-
tion rate of the library is proportional to the effective cycle
numbers of the PCR. The most convenient way to control the
cycle number is to change the amount of target DNA in the
PCR. In theory, 10-fold decrease in template in error-prone
PCR result in log210 ~ca. 3.3-fold elevation in mutation rate,
given that final yield of PCR stay the same.
8. With a high amplification yield of error-prone PCR (>1,000-
fold), one can keep the library free from contaminating the
unamplified wild-type sequence.
9. Mutation rates of the library can be controlled by the concen-
tration of Mn2+ in a PCR reaction. Typical concentration ranges
for Mn2+ is from 0 to 100 μM. Higher concentration results in
the higher mutation rate of the resultant library. In the high
side (>200 μM), polymerase reaction is inhibited, resulting in
the low amount or no product.
10. This step is important to remove the template plasmid. In
addition to this size-fractionation, one can further eliminate
the template plasmid by the treatment of PCR product with
DpnI, which selectively digests plasmid-borne methylated
DNA, in prior to the gel purification.
11. Addition of 10 mM ATP in the reaction mixture would increase
the ligation efficiency.
12. After purification, ligation can be stored at −20°C without
losing the transforming efficiency (library size). In case where
252 M. Furubayashi and D. Umeno

the larger library is necessary, transform the (concentrated)

ligation into commercial super-competent cells and propagate
the transformant in a single mixed culture for 12 h. Miniprep
this mixed culture to prepare plasmid library ready for
transforming to screening strains.
13. Unlike agar surface, nitrocellulose membrane is too smooth to
catch the cell well by using spreader. This way, most of the
transformant are wiped out to the edge of the plate, forming
the colonies only on the perimeter area of the membrane.
Instead, directly spill the diluted transformant culture (1 mL/
plate) on the screening plate. Make sure the media covers the
entire surface of the plate. Wait until the surface gets dried and
then put them into the incubator (37°C).
14. For the better color development, limit the colony density to
be less than 500 colonies/plate by adjusting the culture to be
spilled on each plate.
15. The time needed for colony formation depends both on the
type of E. coli strain and the type of plasmids to be transformed.
XL1-Blue cells harboring two plasmids in Fig. 2 take 20–24 h
to form a colony at 37°C.
16. Two-step incubation is necessary because pigment production
(color development) of C40 carotenoids is very low at 37°C.
On the contrary, incubation/shaking at 37°C is desirable for
growth and plasmid collection.
17. Count and record the number of the colonies formed on the
plate. The total number of the colonies is referred to as “screen-
ing size.”
18. In properly designed screening, clones of interest can be easily
identified by naked eye. For the quantitative ranking and anal-
ysis of the performance of each of the entire library, consider
the use of digital image analysis (16).
19. You might find nothing that appears positive. If that is the
case, it could indicate (1) quality or size of the library was not
high enough (failure in library creation), (2) screening was not
good enough to visualize the activity (failure in screening
design), or simply, (3) the enzyme was not capable of acquiring
the activity by a couple of amino acid substitution (goal too
ambitious).
20. In this system, no IPTG-induction is necessary; the leaky
expression from lac promoter stably supports the expression
needed for cellular pigmentation. Too much expression of car-
otenoid genes is sometimes toxic or even lethal. For instance,
CrtE slows the cell growth, while CrtI is almost lethal upon
full induction.
 14 Directed Evolution of Carotenoid Synthases…
References
1. Lee PC, Momen AZ, Mijts BN, Schmidt-
Dannert C (2003) Biosynthesis of structurally
novel carotenoids in Escherichia coli. Chem Biol
10:453–462
2. Umeno D, Tobias AV, Arnold FH (2005)
Diversifying carotenoid biosynthetic pathways
by directed evolution. Microbiol Mol Biol Rev
69:51–78
3. Schmidt-Dannert C, Umeno D, Arnold FH
(2000) Molecular breeding of carotenoid
biosynthetic pathways. Nat Biotechnol 18:
750–753
4. Wang CW, Liao JC (2001) Alteration of product
specificity of Rhodobacter sphaeroides phytoene
desaturase by directed evolution. J Biol Chem
276:41161–41164
5. Cunningham FX Jr, Gantt E (2001) One ring
or two? Determination of ring number in caro-
tenoids by lycopene epsilon-cyclases. Proc Natl
Acad Sci U S A 98:2905–2910
6. Sun Z, Gantt E, Cunningham FX Jr (1996)
Cloning and functional analysis of the beta-
carotene hydroxylase of Arabidopsis thaliana.
J Biol Chem 271:24349–24352
7. Umeno D, Arnold FH (2004) Evolution of
a pathway to novel long-chain carotenoids.
J Bacteriol 186:1531–1536
8. Schmidt-Dannert C, Lee PC, Mijts BN (2006)
Creating carotenoid diversity in E. coli cells
using combinatorial and directed evolution
strategies. Phytochem Rev 5:67–74
253
9. Sambrook J, Russell DW (2001) Molecular
cloning: a laboratory manual, 3rd edn. Cold
Spring Harbor Laboratory, Cold Spring Harbor
10. Arnold FH, Georgiou G (2003) Directed
evolution library creation: methods and proto-
cols. Humana, Totowa
11. Cirino PC, Mayer KM, Umeno D (2003)
Generating mutant libraries using error-prone
PCR. Methods Mol Biol 231:3–9
12. Umeno D, Tobias AV, Arnold FH (2002)
Evolution of the C30 carotenoid synthase CrtM
for function in a C40 pathway. J Bacteriol
184:6690–6699
13. Umeno D, Hiraga K, Arnold FH (2003)
Method to protect a targeted amino acid resi-
due during random mutagenesis. Nucleic Acids
Res 31:e91
14. Ohnuma S, Nakazawa T, Hemmi H, Hallberg
AM, Koyama T, Ogura K et al (1996)
Conversion from farnesyl diphosphate synthase
to geranylgeranyl diphosphate synthase by ran-
dom chemical mutagenesis. J Biol Chem
271:10087–10095
15. Tobias AV, Arnold FH (2006) Biosynthesis of
novel carotenoid families based on unnatural
carbon backbones: a model for diversification
of natural product pathways. Biochim Biophys
Acta 1761:235–246
16. Tobias AV, Joern JM (2003) Solid-phase screen-
ing using digital image analysis. Methods Mol
Biol 230:109–115